Manual reading of cervical smear slides is better than automation-assisted reading at detecting abnormal cells, finds research published in Health Technology Assessment this week.

Cervical screening currently relies on manually reading the slides. Technology is available which can assist this process by automatically detecting abnormal fields on a slide and presenting them on an automated microscope. Previous research has suggested this could potentially achieve greater sensitivity and productivity.

Researchers, led by Professor Henry Charles Kitchener of the University of Manchester, compared the current manual technique for reading cervical smear slides against the automated technology. Over 70,000 samples were randomised and assessed either by both automated and manual reading or, manual reading only.

The results found that automation was eight per cent less sensitive than manual reading in the detection of abnormal cells. This was largely due to automation-assisted reading failing to detect cases of low grade abnormalities which were detected in manual reading.

“Although automated reading could achieve productivity gains in terms of the numbers of slide checked, on the basis of this evidence there does not appear to be sufficient grounds to recommend automation,” say Professor Kitchener. “The trial also found that the cytoscreeners preferred manual reading as automation-assisted reading was monotonous.”

The full results of this trial have published in Health Technology Assessment, 2011; Vol 15:03

Notes 1. The National Institute for Health Research Health Technology Assessment (NIHR HTA) programme commissions research about the effectiveness, costs, and broader impact of health technologies for those who use, manage and provide care in the NHS. It is the largest NIHR programme and publishes the results of its research in the Health Technology Assessment journal, with over 530 issues published to date. The journal’s 2009 Impact Factor (6.91) ranked it in the top 10% of medical and health-related journals.

Source: National Institute for Health Research